Whose histories are findable? A diversity audit of the Glenbow Archives at UCalgary

Abstract

The Glenbow library and archives at the University of Calgary represent one of Western Canada’s most significant collections of documentary heritage, yet like many large archival repositories, their contents reflect historical patterns of collection and description that privilege mainstream narratives while minimizing others. As stewards of these materials, we question whose histories are represented, and whose are omitted? This paper details a comprehensive collection diversity audit project designed to address that question across Glenbow archival holdings and corresponding digitized materials. The project combines traditional descriptive analysis with distant-reading methods to characterize representation, identify systemic gaps, and surface hidden diversity within archival description. Using only publicly available description, the audit assesses what our users can and cannot discover. We employ topic modelling, sentiment analysis, and visualization across locally defined diversity dimensions. These approaches were used alongside close reading of descriptive records with archivist and librarian knowledge to validate findings. Artificial-intelligence tools, including large language models were incorporated to support metadata analysis, and detect under-described materials across tens of thousands of records. This project aims to offer a replicable model for applying computational and AI-assisted methods to measure diversity dimensions in cultural heritage descriptions. This type of work is essential for providing evidence-based guidance for future acquisitions, digitization priorities, and metadata remediation. We aim to implement a model that can be repeated, applied, and expanded across our collections, and the broader field of archives, libraries, galleries, and museums.